Pembahasan tentang hipotesis dan asumsi penelitian ini disajikan dalam bentuk
tanya jawab. Pertanyaan-pertanyaan yang diajukan dalam pembahasan ini berdasarkan
pengalaman penulis yang sering menerima pertanyaan mahasiswa dan sering melihat
praktek yang salah. Pertanyaan tersebut meliputi pengertian hipotesis dan macam
hipotesis, perumusan hipotesis, perbedaan asumsi dan hipotesis, penyajian hipotesis
teoritis dan hipotesis null dalam thesis atau laporan penelitian, hubungan antara
hipotesis teoritis dan hipotesis statistik, dan jumlah hipotesis alternatif.
1. Apa pengertian hipotesis? Ada berapa macam hipotesis?
Dalam penelitian ada tahapan yang harus dilalui, salah satunya adalah tahapan
hipotesis. Dalam penelitian yang menggunakan rancangan kuantitatif yang melibatkan
beberapa variabel untuk diukur hubungannya (baik hubungan kausal atau hubungan
korelasional) tahapan hipotesis harus dilalui untuk memprediksi secara teoritis hasil
penelitian yang akan dilakukan. Hipotesis ini dilakukan untuk menjawab pertanyaan
penelitian sebelum penelitiannya dilaksanakan. Hipotesis ini disebut hipotesis teoritis.
Dalam proses penelitian kuantitatif yang mengukur hubungan antar variabel,
setelah data terkumpul ada juga tahapan hipotesis yang dilakukan sebelum pelaksanaan
analisis data atau langkah awal dalam proses analisis data. Bahkan analisisnya yang
menggunakan formula statistik bertujuan untuk menguji hipotesis. Hipotesis yang diuji
melalui analisis statistik ini disebut dengan hipotesis statistik.
Lain lagi dalam proses penelitian kualitatif, yang model analisisnya
menggunakan beberapa siklus tahapan, tahap setiap siklusnya melibatkan pengumpulan
data, yang dilanjutkan dengan analisis data, dan berakhir dengan kesimpulan.
Kesimpulan yang diperoleh dari satu siklus ini akan divalidasi dengan siklus berikutnya
untuk memperoleh kesimpulan yang lebih terpercaya, dan diteruskan lagi ke siklus
seterusnya (snow balling technique) sampai dihasilkan kesimpulan yang paling valid.
Sebelum menjadi kesimpulan akhir, kesimpulan dari setiap siklus antara tersebut
3
menjadi kesimpulan sementara (temporary). Kesmipulan sementara ini disebut
hipoptesis empiris.
Lain lagi dalam Penelitian Tindakan Kelas yang bertujuan menghasilkan strategi
pembelajaran inofatif untuk meningkatkan kualitas pembelajaran di kelas. Sebuah
strategi yang dipilih untuk memecahkan masalah pembelajaran yang akan dikerjakan
melalui PTK sudah dipilih sedemikian rupa sehingga peneliti harus yakin dan mampu
memberikankan argumentasi bahwa strategi tersebut akan bisa menyelesaikan masalah
yang dicoba selesaikan. Keyakinan akan tepatnya pilihan strategi ini oleh beberapa ahli
disebut dengan hipotesis tindakan.
2. Bagaimana merumuskan hipotesis teoritis?
Karena hipotesis adalah jawaban teoritis terhadap pertanyaan penelitian, maka
hipotesis dirumuskan berdasarkan rumusan masalahnya, dengan menggunakan bahasa
yang operasional tanpa menggunakan istilah teknis sebagaimana yang biasanya dipakai
dalam perumusan tujuan. Misalnya dalam penelitian yang bertujuan untuk menguji
pengaruh (ini contoh istilah teknis penelitian) jenis kelamin terhadap prestasi belajar
Bahasa Inggris bagi siswa SMA se Malang Raya, masalahnya dirumuskan dengan
kalimat, Apakah prestasi belajar Bahasa Inggris kelompok siswi (perempuan) SMA se
Malang Raya lebih tinggi dibanding prestasi belajar Bahasa Inggris kelompok siswa
(laki-laki) SMA se Malang Raya? Hipotesis untuk rumusan masalah tersebut adalah
prestasi belajar Bahasa Inggris kelompok siswi (perempuan) SMA se Malang Raya
lebih tinggi dibanding prestasi belajar Bahasa Inggris kelompok siswa (laki-laki) SMA
se Malang Raya.
Dalam penelitian yang bertujuan menguji efektifitas (kegiatan yang dilakukan
dengan cara yang sangat teknis, yaitu menggunakan formula statistik) strategi
pembelajaran Listening melalui Lab bahasa, masalahnya dirumuskan dengan kalimat,
Apakah kelompok siswa kelas IX SMPN 1 Malang tahun 2009 yang belajar Listening
melalui lab bahasa berprestasi lebih tinggi dibanding kelompok siswa yang sama yang
tidak belajar Listening melalui Lab bahasa? Hipotesis untuk rumusan masalah tersebut
adalah kelompok siswa kelas IX SMPN 1 Malang tahun 2009 yang belajar Listening
4
melalui lab bahasa berprestasi lebih tinggi dibanding kelompok siswa yang sama yang
tidak belajar Listening melalui Lab bahasa.
Dalam penelitian yang bertujuan untuk mengukur hubungan korelasional (istilah
yang sangat teknis) antara kemampuan membaca dan kemampuan menulis mahasiswa
tingkat pertama jurusan bahasa Inggris Universitas Negeri Malang tahun 2008, rumusan
masalahnya dirumuskan dengan kalimaf, Apakah semakin tinggi kemampuan membaca
mahasiswa tingkat pertama jurusan Bahasa Inggris Universitas Negeri Malang tahun
2008, semakin tinggi pula kemampuan menulis mereka?. Hipotesis untuk masalah ini
dirumuskan dengan kalimat Semakin tinggi kemampuan membaca mahasiswa tingkat
pertama jurusan Bahasa Inggris Universitas Negeri Malang tahun 2008, semakin tinggi
pula kemampuan menulis mereka.
3. Apa beda asumsi dan hipotesis?
Setiap saat orang melakukan suatu kegiatan, pasti dia memiliki keyakinan
terhadap apa yang dia lakukan. Atas dasar keyakinan itulah dia melakukan kegiatan
tersebut. Bila keyakinan tersebut tidak ada, dia tidak mungkin melakukan kegiatan
tersebut. Seorang yang sedang menyeberang jalan (setelah menengok ke kanan dan kiri)
dia yakin bahwa dia akan aman untuk menyeberang jalan. Kalau dia tidak yakin bahwa
dia akan aman untuk menyeberang jalan, dia tidak mungkin menyeberang jalan itu.
Mahasiswa yang belajar di S2 UM yang bertujuan untuk meningkatkan kemampuan
akademiknya (selain untuk memperoleh ijasah S2 untuk kepentingan sertifikasi,
misalnya) punya keyakinan bahwa melalui kegiatan belajar di S2 UM dia akan bisa
berhasil meningkatkan kemampuan akademiknya dan bisa lulus dengan membawa
isajasah S2 UM dua tahun sejak dia memulai semester pertama. Kalau mahasiswa
tersebut tidak yakin bisa meningkatkan kemampuan akademik yang dia perlukan dan
tidak yakin bisa lulus dengan membawa ijasahnya, dia tidak mungkin masuk ke
program S2 UM. Keyakinan itu disebut asumsi. Jadi asumsi adalah keyakinan yang
dimiliki seseorang sebagai syarat orang tersebut memutuskan untuk melakukan suatu
kegiatan. Tanpa asumsi seseorang tidak akan memutuskan untuk melakukan sesuatu
perbuatan.
5
Orang Islam meyakini atas dasar ajaran yang diikutinya, bahwa segala sesuatu
yang ada di alam ini (termasuk dirinya) diciptakan oleh Tuhan lengkap dengan cara
kerjanya masing-masing. Apabila sistem yang ada pada sebuah objek itu berjalannya
terganggu karena sesuatu hal (apabila sistem yang harus berjalan dalam tubuhnya
terganggu), maka objek tersebut tidak bisa berfungsi dengan baik (apabila yang
terganggu adalah sistem kerja tubuhnya, maka tubuhnya akan sakit). Keyakinan itulah
yang dalam penelitian disebut asumsi.
Dalam kegiatan penelitian, peneliti yang bertujuan untuk menemukan aturan
(sistem, pola, atau formula) yang menjadi hukum berfungsinya objek penelitian,
memiliki asumsi bahwa objek penelitian di alam ini memiliki sistem (pola, aturan, atau
formula) kerja tertentu dan sistem kerja itu bisa ditemukan melalui penelitian. Ary,
Jacobs, dan Razafieh (1979:13) menyebutkan beberapa pernyataan bahwa
a fundamental assumption made by scientists is that the events they investigate
are lawful or ordered no event is capricious.
This assumption underlies any statement that declares that under specific
conditions certain events will occur.
Related to this first assumption is the belief that the events in nature are orderly
and regular and that this order and regularity of nature can be discovered
through the scientific method.
Berangkat dari asumsi adanya sistem pada objek penelitian yang bisa ditemukan
melalui penelitian itulah peneliti berusaha menemukan sistem tersebut melalui kegaiatn
penelitian. Tidak mungkin peneliti berusaha menemukan sistem tersebut kalau dia tidak
yakin bahwa sistem itu ada atau dia yakin sistem itu ada tetapi tidak yakin bisa
ditemukan melalui penelitian.
Asumsi tidak untuk diuji tetapi sudah otomatis menjadi dasar tindakan, sementara
hipotesis adalah prediksi terhadap apa yang akan dihasilkan dari penelitian yang
dilakukan.peneliti.
6
4. Haruskah hipotesis teoritis dan hipotesis null disajikan bersama pada chapter 1
dalam tesis atau laporan penelitian?
Dalam chapter 1 pada tesis atau laporan penelitian disampaikan latar belakang
yang menjadi titik tolak mengapa satu penelitian layak untuk dilaksanakan, pertanyaan
yang dirumuskan (research problems) untuk dijawab melalui penelitian yang
dikerjakan, tujuan (research objectives) yang ingin dicapai melalui kegiatan penelitian,
asumsi yang mendasari pertanyaan penelitian, hipotesis teoritis (theoritical hypothesis)
yang memprediksi hasil penelitian yang dikerjakan, kontribusi hasil penelitian terhadap
pengembangan teori (theoretical significance) dan atau terhadap praktek pembelajaran
(practical significance), penjelasan kata-kata kunci (key terms) yang digunakan dalam
proses kegiatan penelitian, dan ada juga yang dilengkapi dengan uraian singkat tentang
kerangka teori (theoretical framework) yang mendasari cara kerja kegiatan penelitian
yang dikerjakan.
Dari uraian di atas nampak tidak ada tempat untuk membahas topik apapun di
chapter 1 tesis atau laporan penelitian selain topik-topik tersebut. Penempatan hipotesis
null pada chapter 1 (sebagaimana banyak ditemukan dalam tesis S2 dan banyak laporan
penelitian) adalah salah tempat karena hipotesis null digunakan dalam proses analisis
data yang menjadi bagian dari metode penelitian yang disajikan pada chapter 3.
5. Apa hubungan antara hipotesis teoritis dan hipotesis statistik?
Hipotesis teoritis yang benar bisa diuji secara statistik. Hipotesis yang bisa diuji
secara statistik ini selalu melibatkan lebih dari satu variabel. Hubungan antar variabel
(kausal atau korelasional) inilah yang akan diuji secara statistik. Tentunya pengujian
hipotesis teoritis ini tidak dilakukan secara langsung, tapi harus mengikuti prosedur
pengujian hipotesis, yang menuntut beberapa langkah. Langkah pertama adalah
melakukan transformasi hipotesis teoritis menjadi hipotesis null (disingkat Ho). Dengan
kata lain apabila hipotesis teoritis tidak bisa ditransformasikan ke dalam Ho, maka
hipotesis teoritis tersebut salah.
Misalnya: Karena semua mahasiswa baru program Studi S2 Pendidikan Bahasa
Inggris (PSPBI) Universitas Negeri Malang (UM) diterima melalui seleksi yang ketat
yang menjamin kompetensi yang memadahi untuk menyelesaikan studi di S2 PSPBI
7
UM, maka bisa diprediksikan bahwa mereka semua akan lulus tepat waktu. Pernyataan
tersebut secara logika benar dan berbunyi seperti hipotesis teoritis, tetapi pernyataan
tersebut bukan hipotesis teoritis, karena tidak berisi lebih dari satu variabel sehingga
tidak bisa ditransformasikan ke dalam hipotesis null dan oleh karena itu tidak bisa diuji
secara statistik.
Contoh Hipotesis teoritis:
(1) Prestasi dalam belajar listening kelompok mahasiswi semester I jurusan
bahasa Inggris UM cenderung lebih tinggi dibanding prestasi kelompok
mahasiswa pada semester I jurusan yang sama.
(2) Kelas siswa SMPN 1 Malang yang belajar listening dalam lab bahasa
cenderung memiliki hasil kemampuan listening yang lebih tinggi dibanding
dengan kelas siswa di SMP yang sama yang tidak belajar dalam lab
bahasa.
(3) Semakin tinggi penguasaan vocabulary siswa-siswa SMA di Malang
cenderung semakin tinggi pula kemampuan reading comprehensionnya.
Hipotesis teoritis tersebut bisa ditransforamsikan ke dalam hipotesis null menjadi:
(1) Prestasi dalam belajar listening kelompok mahasiswi semester I jurusan
bahasa Inggris UM cenderung tidak berbeda dengan prestasi kelompok
mahasiswa pada semester I jurusan yang sama
(2) Kelas siswa SMPN 1 Malang yang belajar listening dalam lab bahasa
cenderung memiliki hasil kemampuan listening yang tidak berbeda
dibanding dengan kelas siswa di SMP yang sama yang tidak belajar dalam
lab bahasa.
(3) Semakin tinggi penguasaan vocabulary siswa-siswa SMA di Malang tidak
diikuti dengan semakin tingginya kemampuan reading comprehensionnya
Langkah ke dua adalah menganalisis data untuk melihat apakah ada bukti cukup
(signifikan) untuk menolak hipotesis null. Hasil analisis bisa menunjukkan tidak ada
bukti signifikan untuk menolak hipotesis null atau ada bukti signifikan untuk
menolak hipotesis null. Penentuan tingkat signifikansi bukti empiris (dari data)
dilakukan atas dasar penghitungan statistik. Apabila hasil analisis menunjukkan tidak
ada bukti signifikan untuk menolak hipotesis null, maka kesimpulan akhir penelitian
telah bisa dirumuskan, yaitu fakta (data) empiris tidak (cukup untuk) menolak
pernyataan hipotesis null. Dengan tidak adanya bukti signifikan untuk menolak
8
hipotesis null, maka hipotesis teoritisnya hanya dinyatakan benar dalam teori, tetapi
tidak didukung oleh data empiris.
Apabila hasil analisis menunjukkan adanya bukti signifikan untuk menolak
hipotesis null, maka pernyataan hipotesis null (seperti pada contoh di atas: Prestasi
kelompok mahasiswi jurusan Bahasa Inggris UM cenderung tidak berbeda dengan
prestasi kelompok mahasiswa pada jurusan yang sama) ditolak oleh kenyataan empiris.
Penolakan hipptesis null oleh bukti data signifikan menghasilkan kesimpulan sementara
yang mengubah hipotesis null tersebut (dalam contoh tersebut hipotesis null berubah
menjadi Prestasi kelompok mahasiswi jurusan Bahasa Inggris UM cenderung berbeda
dengan prestasi kelompok mahasiswa pada jurusan yang sama). Pernyataan berbeda
dalam contoh hipotesis tersebut belum mengindikasikan kelompok mana yang lebih
tinggi. Untuk itu diperlukan langkah ketiga dalam proses analisis ini, yaitu penentuan
hipotesis alternatif dan folow-up tes untuk meentukan kelompok mana yang lebih
tinggi.
6. Ada berapa hipotesis alternatif?
Hipotesis null bisa diikuti oleh satu hipotesis alternatif atau dua (atau lebih)
alternatif hipoptesis tergantung pada implikasi atau tujuan penelitiannya. Penelitian
yang bertujuan menguji efektifitas pembelajaran listening melalui lab bahasa akan
memiliki implikasi terhadap kebijakan sekolah untuk membeli lab bahasa atau tidak
perlu membeli lab bahasa. Karena lab bahasa bukan sesuatu yang murah, maka sebelum
memutuskan pembelian lab bahasa, sekolah perlu mendapat informasi yang meyakinkan
dari hasil penelitian apakah memang penggunaan lab bahasa lebih efektif untuk
pembelajaran listening dibanding dengan pembelajaran listening melalui kelas
konvensional seperti yang telah rutin dikerjakan di sekolah tersebut. Dengan demikian
penelitian tersebut sepenuhnya diarahkan untuk melihat apakah kelompok siswa yang
belajar listening menggunakan lab bahasa memang berhasil jauh (signifikan) lebih
tinggi dibanding dengan kelompok siswa yang belajar Listening dalam kelas
konvensional rutin seperti biasa. Arah lainnya, apakah kelompok siswa yang belajar
Listening melalui kelas konvensional berprestasi lebih tinggi dibanding dengan
kelompok siswa yang belajar Listening menggunakan lab bahasa, tidak menjadi
kepedulian peneliti karena jawaban ini tidak memiliki kepentingan (tidak akan
9
mempengaruhi kebijakan) sekolah. Dalam kontek seperti ini penelitian tersebut hanya
memiliki satu alternatif hipotesis, yang biasanya dirumuskan seperti hipotesis teori.
Sementara penelitian yang bertujuan menguji pengaruh jenis kelamin terhadap
prestasi belajar Bahasa Inggris semua siswa SMP se Malang Raya memiliki arah ganda.
Bila ditemukan cukup (signifikan) bukti untuk menolak hipotesis null (prestasi
kelompok siswi tidak berbeda dengan prestasi kelompok siswa SMP se Malang Raya
dalam belajar bahasa Inggris), maka perlu dilakukan follow-up uji hipotesis berikutnya,
yaitu hipotesis alternatif satu yang perumusannya seperti pada hipotesis teoritis
(prestasi kelompok siswi lebih tinggi dibanding prestasi kelompok siswa SMP se
Malang Raya dalam belajar bahasa Inggris) dan hipotesis alternatif dua yang
perumusannya kebalikan dari rumusan hipotesis teoritis (prestasi kelompok siswa
berprestasi lebih tinggi dibanding dengan prestasi kelompok siswi SMP se Malang
Raya dalam belajar bahasa Inggris). Kedua arah ini sama-sama ingin diketahui oleh
peneliti, yaitu apakah prestasi siswa lebih tinggi atau prestasi siswi lebih tinggi. Dalam
konteks seperti ini, penelitiannya menggunakan dua hipotesis alternatif (dua arah).
Seorang peneliti yang juga kebetulan seorang guru baru yang akan mengajar ingin
mengetahui mana di antara dua strategi yang telah dikuasai lebih efektif yang akan
dipilihnya dalam mengajar. Penelitian yang dia lakukan akan melibatkan dua alternatif
hipotesis. Tetapi seorang guru lama yang telah bertahun-tahun menggunakan satu
strategi pembelajaran, apabila dia mengenali satu strategi pembelajaran baru yang
secara teoritis lebih efektif, dia tidak langsung mengganti strategi lamanya dengan
strategi baru. Dia perlu mendasarkan pilihanya pada hasil penelitian. Apabila dia
melakukan penelitian yang membandingkan efektifitas strategi lama dengan strategi
baru tersebut, maka dia hanya akan menggunakan satu alternatif hipotesis.
KEPUSTAKAAN
Ary, D., Jacobs, L.C., Razavieh, A. 1979. Introduction to Research in Education.
Chicago: Holt, Rinehart and Winston, Inc.
10
Hypothesis in Language Learning Research
By Mohammad Adnan Latief
Abstract: Hypothesis is very often inevitable in research activities.
Hypothesis is of at least three kinds, each of which should not be
confused. A study trying to measure the relationship between
variables can predict the finding based on theory or logical common
sense. This prediction is called theoretical hypothesis. In testing
hypothesis quantitatively, the theorretical hypothesis should be
transformed into statistical hypothesis, which takes the form of Null
hypothesis and its alternatives. It is the Null hypothesis that is to be
tested to justify its rejection or otherwise its acceptance. In
qualitative study, the result of first data analysis is called temporal
empirical hypothesis that should be validated with more data. This
cycle of rechecking the result with more data is done again and again
until the hypothesis becomes the final conclusion.
Key words: theoretical hypothesis, empirical hypothesis, and
statistical hypothesis.
Drawing a research hypothesis is very often inevitable in a research project
involving either quantitative or qualitative data. The type of hypopthesis is different from
one research activity to another research activity, or from research involving quantitative
data from the one involving qualitative data. In other words, there are more than one
types of hypothesis each of which has its definition and function that should not be
confused by the researcher. Misplacing one hypothesis from another hypothesis shows
that the researcher does not have clear understanding of what hypotheis actually is.
From my experience reading graduate theses and attending graduate students
seminar clases I see a lot of misplaced hypotheses. Very often graduate students confuse
one kind of hypothesis from another kind of hypothesis. They confuse theoretical
hypothesis with statistical hypothesis, hypothesis for quantitative research with one for
qualitative research. This article tries to make clarification for language researchers
including graduate students doing research for their thesis so that they can choose the
right kind of hypothesis in their research report for their thesis.
11
When a researcher predicts the finding of his/her research based on a theory or
based on a logical common sense, he is drawing a theoretical hypothesis. Another
researcher who draws a temporary conclusion, a conclusion that he/she will validate with
more data, is drawing an empirical hypothesis. While a researcher is computing data for
his/her research, he wants to be objective, he does not want to be biased. Therefore,
he/she will have to state a null hypothesis with its alternative(s). This kind of hypotheis
is called statistical hypothesis, different from the other two hypotheses.
THEORETICAL HYPOTHESIS
In a research project applying quantitative approach or which relies the expected
finding on quantitative data (data that have to be quantified in the process of analysis), a
researcher who is trying to measure the relationship between two or more variables must
predict the answer to the problem or the finding of the research based on theory or based
on logical common sense. This prediction must have theoretical as well as logical truth
in it. This theoretical and logical prediction is called hypothesis (Tuckman, 1999:89;
Borg & Gall, 1983:78). More precisely, Ary, Jacobs, and Razavieh (1979:72) defines
the theoretical and logical hypothesis as
a tentative proposition suggested as a solution to a problem or as an
explanation of some phenomena. It presents in simple form a statement of
the researcher s expectation relative to a relationship between variables
within the problem .
For example, in a research aimed at testing the effect of sex toward the achiemenet
in language learning or in math learning, the researcher can predict that theoretically and
logically female students would make better achievement in language learning than male
students of the same level. In the same way he/she can predicat that theoretically and
logically, male students would make better achievemnt in math learning than female
students of the same level.
When a study does not try to measure the relationship between two or more
variables, or when it involves only one variable, a hypothesis cannot be provided. In
other words, not every research problem has to be followed with a research hypothesis.
12
Forcing one s self to provide hypothesis for every research problem would result in the
difficulties during the process of statistical computation of the data, as every hypothesis
should later be tramsformed into a statistical hypothesis for the purpose of testing. So, a
hypothesis which cannot later be transformed into a statistical hypothesis is a wrong
hypothesis (Ary, Jacobs, and Razavieh, 1979: 77). In the same way, a hypothesis which
already takes the form of statistical hypothesis, not for the purpose of testing, is also
wrong because it does not reflect theoretical and logical prediction of the answer to the
research problem.
Tuckman summarizes three criteria to meet in stating a hypothesis correctly. A
good hypothesis statement should
(1) conjecture about the direction of the relationship between two or more
variables, (2) be stated clearly and unambiguously in the form of a
declarative sentence, and (3) be testable; that is, it should allow restatement
in an operational form that can then be evaluated based on data (Tuckman,
1999:73).
Following are examples of good hypothesis statements derived from the
related research problems.
Problem: Do students with higher IQ tend to achieve better in learning a
foreign language than those with lower IQ?
Hypothesis: Students with higher IQ tend to achieve better in learning a
foreign language than those with lower IQ.
Problem: Do students learning a foreign language achieve better from a
directive foreign language teacher than those from a non-
directive foreign language teacher?
Hypothesis: Students learning a foreign language achieve better from a
directive foreign language teacher than those from a non-
directive foreign language teacher.
Problem: Does the repetitious use of prompting in instructional
language learning materials impair the effectiveness of those
materials?
Hypothesis: Students using language learning materials with repetitious
use of prompting do not learn as effectively as those using
language learning materials without repetitious use of
prompting. (Tuckman, 1999: 29, 73)
STATISTICAL HYPOTHESIS
13
Unlike theoretical hypothesis aimed at predicting the outcome of the research,
statistical hypothesis is a statement that has to be made by the researcher while
computing the data to allow statistical hypothesis testing (Tuckman, 1999: 88). This
statistical hypothesis is transformed from the theoretical hypothesis and should, therefore,
not be mentioned after the statement of the research problem in Chapter 1 of the research
report or thesis. In fact, it does not have to be reported in any chapter of the research
report or thesis. It is usually found in Statistics books which describe the process of
computation, more specifically in the process of hypothesis testing. Statistical hypothesis
always takes the form of null hypothesis, represented as Ho, followed with its
alternative(s), alternative one hypothesis, represented as H1, and alternative two
hypothesis, represented as H2. It is this Ho that is to be tested in the statistical analysis.
For example, from the theoretical hypothesis mentioned earlier about the effect of
sex toward achievement in language learning, the hypothesis which states that female
students would make better achievement in language learning than male students of the
same level can be transformed into a statistical Ho statement, there is no difference in
achievement between female students and male students in language learning
(represented as Mean Score of Female students - Mean score of male students= 0). From
this Ho, the alternative H1 can be formulated as Mean score of female students is
significantly bigger than that of male students, and its alternative H2 as Mean score of
male students is significantly bigger than that of female students.
Statistical analysis tests whether there is enough statistical evidence to reject Ho.
If enough evidence is found, then the researcher rejects the Ho. In the absence of enough
statistical evidence, the researcher accepts the Ho. In other words, statistical analysis
allows the researcher to test Ho and to determine whether the evidence suggets rejecting
or accepting it. If the statistical analysis results justify rejecting the Ho, then they provide
support for its alternative hypothesis (Tuckman, 1999: 284). If the statistical analysis
results do not justify rejecting Ho, then Ho is accepted, which means that there is no
support for its alternative hypothesis.
14
In the case of testing the effect of sex towards achievement in language learning,
Ho would refer to the belief that the average score of male students is about the same as
that of female students. Even if the average score is different, let s say the mean score of
the female students is higher than that of male students, it is believed that the difference
occurs by mere chance variations. If the measurement is to be repeated, the difference in
the average score might disappear. Testing Ho refers to testing that believe.
If the statistical hypothesis testing results do not show strong evidence to reject that
belief (Ho), then that belief (Ho) is not to be changed and is then taken as the finding.
That belief of equality between the two means (Ho) will change only if the statistical
hypothesis testing results show strong evidence to reject it. Rejecting Ho means that
there is a strong reason to believe that the difference is not mere chance variations, it is a
real difference (Vockell & Asher, 1995:316-319). This means support to the alternative
hypothesis.
EMPIRICAL HYPOTHESIS
A language learning researcher ((as illustrated by Bogdan and Biklen, (1998:64) in
their hypothetical study for descriptive illustration)) is trying to study effective language
teachers. His general topic and focus is developing a theory on effective language
teachers. He starts his research by deciding to choose a teacher who can represent a real
effective language teacher as his subject. After observing several language classrooms
and interviewing several experienced language teachers, language learners, school
headmasters, and some parents whose children are learning language in their classrooms,
he gets recommendation to select one teacher who is believed to have the right authority
to represent or meets the expected criteria of an effective language teacher as his subject
of the study.
He then conducts an in-depth interview in a long, open-ended, tape recorded
discussion with the selected effective language teacher. As a supplement to the
interview, he also visits the teacher s schools and observes her in action. From that initial
interview and observation, he develops a loose descriptive theory of language teacher
15
effectiveness. This theory is his first conclusion of the study which is still temporary as
it is based on initial interview and observation. This is his first temporary empirical
hypothesis.
After he has sketched out his temporary theory of effective language teacher, he
picks a second effective language teacher to interview. In picking the following
effectieve language teachers to interview, he used snowball sampling technique; that is,
he asked the first person he interviewed to recommend others. He interviewed the second
subject in a similar open-ended manner, witholding the theory (first Hypothesis) he
developed from his first initial interview. After the second interview, he rewrites and
modifies the theory (the first hypopthesis) to fit the new case. He continues choosing and
interviewing new people, modifying the theory (previous hypothesis) to fit the new case.
He proceeds in this manner, picking new subjects, enlarging the theory (the temporary
hypothesis) until no longer comes across any case that does not fit the theory (the
hypothesis). At the end, he takes the final hypothesis as the conclusion or the finding of
his study about language teacher effectiveness (Bogdan & Biklen, 1999:64).
The narration on the process of qualitative research on language teacher
effectiveness above shows us another type of hypothesis (Tuckman, 1999: 417) different
from the earlier mentioned theoretical as well as statistical hypothesis. This type of
hypotheis is not based on theory, nor is it for the purpose of statistical data computation,
it is a temporary hypothesis based on data analysis. This hypothesis after going through
several modification to fit several cases becomes the final conclusion of the study.
Following is another illustration to clarify empirical hypothesis. A study on the
order of acquisition on English reflexive pronouns among Spanish children learning
English as a Second Language in the United States reveals that the children first acquire
the form: myself, yourself, herself, ourself*, hisself*, and theirself*. The first three
reflexive pronouns myself, yourself, and herself are correctly learned in this first
temporary conclusion but the second three reflexive pronouns hisself*, ourself*, and
theirself* are overgeneralized from the possessive pronouns my, your, her, his, our and
their that they have learned before plus the word self. Then in the next stage in the
16
sequence of acquisition of the reflexive pronouns, the children learn : myself, yourself,
herself, himself, ourself*, and themself*. This second stage of the sequence in the
acquisition of reflexive pronouns is the improvement of the first stage conclusion,
correcting the form hisself* and theirself* into himself and themself*. This second stage
conclusion is still temporary, it is not final yet. Only in the next stage, then the Spanish
children acquire the final conclusion myself, yourself, herself, ourselves, himself, and
themselves, correcting the wrong forms ourself* and themself* to the correct forms
ourselves and themselves.
The process of acquiring the rules of reflexive pronouns by the Spanish children
learning English as the Second Language in the United States that takes stages shows that
learning the rules of reflexive pronouns for these children does not take once for all. It
starts from the first exposure to the actual use of the expressions involving those reflexive
pronouns, then is followed spontaneously with subconscious analysis by the children
resulting in the first stage conclusion: myself, yourself, herself, ourself*, hisself*, and
theirself*. The first stage conclusion is verified and revised through the second cycle
which again starts with further exposure and ends in subconscious analysis resulting in
the second stage conclusion: myself, yourself, herself, himself, ourself*, and themself*.
The second stage conclusion is again verified and revised through the third cycle which
again starts with further exposure and ends in subconscious analysis resulting in the last
stage conclusion: myself, yourself, herself, himself, ourselves, and themselves.
This process of acquisition of the rules of reflexive pronouns by the Spanish
children learning English as a Second Language in the United States is just like the
process of data analysis in qualitative research that involves a temporary empirical
hypothesis, which is revised into the next temporary empirical hypothesis, until the
hypopthesis becomes the conclusion (Dulay, Burt, & Krashen, 1982:215)
REVIEW ON HYPOTHESIS STATEMENTS IN THESIS
17
Following is discussion on how hypotheses are stated in theses on language
learning. Some theses state good and clear hypotheses, some state wrong hypothesis, and
still others state hypotheses which are not elaborated enough.
Good and Clear Hypotheses
Following are examples of hypothesis statements that meet the criteria of good
hypothesis. They state the expected relationship between variables involved, are testable,
consistent with the existing body of knowledge, and stated simply and concisely (Ary,
Jacobs, and Razavieh, 1979:77-79)
The higher scores of the students knowledge on topics are, the higher
scores of their ability in writing a coherent expository discourse
(Dwimaretno, 1996:5)
There is a significant positive correlation between frequency of watching TV
serials and speaking ability (Ambarkati, 1991:15)
The high scores of students experience in English songs tend to go together
with high scores of students English pronunciation (Sulistyaningtyas,
1996:3)
As class level increases,
the average score for descriptive papers should also increase,
the average coherence score for descriptive papers should also increase,
the average complexity of sentences per paper should also increase,
the average number of sentences per paper should also increase,
the average frequency of grammatical errors per paper should decrease,
the average frequency of mechanical errors per paper should decrease.
(Latief, 1990:6-7).
The average score for argumentative papers of the seniors should be higher
than that of the juniors (Latief, 1990:7).
The average coherence score for argumentative papers of the seniors
should be higher than that of the juniors (Latief, 1990:7).
Students learning vocabulary through distributed practice--practice in
learning vocabulary through several short sessions separated by one or two
days-- achieve better than those learning vocabulary through massed
practice --practice in learning vocabulary through a few long sessions
separated by one week or two (Djiwandono, 2001: 195).
18
Achievement in language learning by K1 students is significantly better than
that by K2 students (Gosong, 1993:20).
Students who are taught with summarizing technique perform better in
literal reading comprehension skill than those taught with non-summarizing
technique (Junaidi, 1996:10).
The mean of vocabulary gain score of the group under individualized
vocabulary instruction is higher than that of the group under teacher-
centered instruction (Masduki, 1999:11).
The total group of learners treated with individualized technique of
instruction have better written performance of the English verb have and
tense auxiliary have than the other total group treated with traditional full-
class teaching (Harsono, 1993:16).
Misplaced Hypothesis
Following are examples of hypothesis statements which are misplaced as they are
null hypothesis instead of theoretical hypothesis which function should be to predict the
answer to the problem of the study based on theory or logical common sense.
There is no difference in reliability between Cloze Test and Multiple Choice
Test which are based on different texts with the same level of readibility in
measuring reading Comprehension ability (Atmoko, 1991:5)
There is no difference between the mean of vocabulary gain score of the
group under individualized vocabulary instruction and the mean of
vocabulary gain score of the group under teacher-centered instruction
(Masduki, 1999:10).
There is no difference in discrimination index between Cloze Test and
Multiple Choice Test which are based on different texts with the same level
of readibility in measuring reading Comprehension ability (Atmoko,
1991:5)
There is no difference in level of difficulty between Cloze Test and Multiple
Choice Test which are based on different texts with the same level of
readibility in measuring reading Comprehension ability (Atmoko, 1991:5)
Not Clearly Stated Hypothesis
19
Some hypothesis statements need elaborations to show the detailed plan of the
research strategy in analysis. Following are examples of the hypothesis statements which
need more elaborations.
There is a significant relationship between the use of syntactic clues and the
use of semantic clues (Kusumarasdyati, 1996:7).
The mean of structure test resulted by English for Beginners is significantly
better than the mean resulted by the Bahasa Inggris 2a untuk SMP
published by Balai Pustaka (Waskito, 1994;10)
The mean of interview test resulted by English for Beginners is
significantly better than the mean resulted by Bahasa Inggris 2a untuk
SMP published by Balai Pustaka (Waskito, 1994;10)
Kusumarasdyati s hypothesis statement which refers to correlational data analysis
technique could be made more operational into:
The more the students are able to identify the syntactic clues the better they
are in identifying the semantic clues.
Waskito s hypothesis statements which refer to comparative data analysis technique
could be made more operational into:
The mean of the Structure test scores for the students who are taught using
English for Beginners is significanly higher than that for the students who
are taught using Bahasa Inggris 2a untuk SMP Published by Balai
pustaka.
The mean of the interview test scores for the students who are taught using
English for Beginners is significanly higher than that for the students who
are taught using Bahasa Inggris 2a untuk SMP Published by Balai
pustaka.
Ambiguosly Stated Hypothesis
Following are examples of hypothesis statements which need more elaboration to
avoid ambiguity in the design of research to be applied.
The students scores on Reading Comprehension test elicited with MCT format are
highly correlated with their scores on Reading Comprehension test elicited with
SAT format (Mahmud, 1996:4)
20
There is a significant difference in students writing achievement of all course
levels as shown by their average scores on the expository writing test (Mukminatin,
1997:13)
There is a significant difference in students writing achievement of all course
levels as shown by their average scores on the narrative writing test (Mukminatin,
1997:13)
There is a significant difference between the group with Bloom s taxonomy of
cognitive levels of questions which put more stress on CAA types of questions and
the group with KCA (traditional) types of questions on students reading ability
(Surjosoeseno, 1991:8)
There would be a significant relationship between students paragraph
knowledge and their ability to write an expository paragraph (Rohimah,
1992:6)
For the students with good English achievement, the frequency of general
and local strategies used in reading an indonesian text is different from
that used in reading an English text (Suharmanto, 2000:16).
For the students with poor English achievement, the frequency of general
and local strategies used in reading an indonesian text is different from
that used in reading an English text (Suharmanto, 2000:16).
For the students with good English achievement and poor English
achievement the frequency of general and local strategies used in reading
an indonesian text is different from that used in reading an English text
(Suharmanto, 2000:16).
The use of Indonesian as daily communication medium in students family
life is related to the achievement of those students in learning Indonesian at
school (Gosong, 1993:20).
There is a difference in the students writing ability
between those who use clustering strategy and those who do not use
mapping strategy, between those who use listing strategy and those who
do not use mapping strategy, between those using outlining strategy and
those using listing strategy, between those using listing strategy and
those using outliningng strategy, between those using outlininging
strategy and those using clustering strategy (Telaumbanua, 1992:16)
The term highly correlated by Mahmud or significant relationship by Rohimah and
by Kusumarasdyati in the above examples may lead to two possible interpretations;
21
negatively corelated or positively correlated. A good hypothesis states clearly whether the
correlation is going to be negative or positive, as follows.
The higher the students scores on Reading Comprehension elicited with
MCT format are, the higher their scores on Redaing Comprehension elicited
with SAT format are. (Positive correlation),
or
The higher the students scores on Reading Comprehension elicited with
MCT format are, the lower their scores on Redaing Comprehension elicited
with SAT format are. (Negative correlation).
The higher the students paragraph knowledge is, the higher their ability in
writing an expository paragrapgh (positive correlation)
or
The higher the students paragraph knowledge is, the lower their ability in
writing an expository paragrapgh (negative correlation).
The statement there is a significant difference by Mukminatin could be made
more specific like the higher the course level is, the better the achievement is. Or the
seniors have better writing achiements than the juniors as shown by their average
scores on expositiry writing test. Suharmanto s statement is different from that of ,
Surjosoeseno s and Telaumbanua s statement There is a difference does not predict
theoretically which one is going to be better. Suharmanto states the hypothesis for good
achievers, for low achievers, and for both the high and low achievers. The last
hypothesis for the high and low achievers does not give clear reference who they are.
Gosong s statement ..is related to.. again gives another example of hypothesis statement
which does not show the direction of the relationship between the variables involved.
Wrong Hypothesis
Following are examples of hypothesis statements which are wrong as they are not
easily transformable to statistical hypothesis.
Learning English at the elementary school age supports the ability of the
students in learning the language at SMP (Ivonne, 1995:4)
22
The word support is not an operational statement for a hypothesis which tries to test
correlational relationship between the variables involved. In fact, the word support is
closer to the interpretation of causal relationship to correlational relationship between
variables. A suggested better statement of hypothesis sould read like this:
The longer and more intensive English course the elemantary school age
students experience, the better they will succeed in learning English at
SLTP.
CLOSING
Hypothesis is of several types, each of which has to be stated according to its own
function and placed accordingly. Several theses have shown ambiguous hypothesis
statements, some have shown wrong hypothesis statements, some do not state hypothesis
clearly, and still some others even misplace their hypothesis. Good, clear, and correct
hypothesis statements are helpful to the researchers themselves as well as to the readers
in guiding where the research is up to.
REFERENCES
Ambarkati, sul. 1991 The Correlation between the Frequency of Watching TV Serials
and Speaking Ability of S1 Students of English Department at IKIP MALANG.
Unpublished undergraduate Thesis, FPBS, IKIP MALANG.
Atmoko, Rudi., M. 1991 The Effectiveness of Cloze Test and Multiple Choice Test in
Measuring Reading Comprehension Ability Based on Difficult Texts. Unpublished
undergraduate Thesis, FPBS, IKIP MALANG.
Ary, Donald., Jacobs, Lucy, C., Razavieh, Asghar. 1979 Introduction to Research in
Education, Second edition. Holt, Rinehart and Winston.
Bogdan, Robert, C., Biklen, Sari, K. 1998 Qualitative Research for Education, An
Introduction to Theory and Methods, Third edition. Boston: Allyn and Bacon.
Borg, Walter, R., Gall, Meredith, D. 1983 Educational Research. An Introduction,
Fourth Edition. White Plains: Longman Inc.
23
Djiwandono, Patrisius, I. 2001. Pengaruh Post-Lesson Quiz terhadap Penguasaan Kata
oleh Pembelajar Bahasa Inggris. Suatu Studi pada Pembelajar di Tingkat
Menengah Lanjut. Jurnal Penelitian Kependidikan, Lembaga Penelitian
Universitas Negeri Malang, Tahun 11 Nomor 2 desember 2001.
Dulay, Heidi., Burt, Marina., Krashen, Stephen. 1982. Language Two. Oxford: Oxford
University Press.
Dwimaretno, Soeraswati. 1996. The Correlation between the Students Knowledge on
Topics and their Ability in Writing a Coherent Expository Discourse. Unpublished
undergraduate Thesis, FPBS, IKIP MALANG.
Gosong, I Made, 1993. Perbandingan Prestasi Belajar Bahasa Indonesia antara Siswa
yang Menggunakan Bahasa Indonesia dan Siswa yang Menggunakan Bahasa Bali
dalam Keluarga, unpublished thesis, Graduate School, IKIP MALANG.
Harsono, Yohanes, M. 1993. The Effect of Restricted Individualized Technique of
Instruction on the Achievement of learners Written Performance of the English
Verb HAVE and Tense Auxiliary HAVE (The case at Atma Jaya university
Jakarta), unpublished thesis, Graduate School, IKIP MALANG.
Ivonne, 1995. The Effect of Learning English in the Elementary School Age on the
Students Achievement in English as a School Subject at COR JESU Malang.
Unpublished undergraduate Thesis, FPBS, IKIP MALANG.
Junaidi, 1996. The Effectiveness of Summarizing as a Technique of Teaching English
Reading Comprehension. unpublished thesis, Graduate School, IKIP MALANG.
Kusumarasdyati, 1996 The Use of Context Clues by the Students of English IKIP
MALANG in Reading Comprehension, unpublished thesis, Graduate School, IKIP
MALANG.
Latief, Mohammad, A. 1990. Assessment of English Writing Skills for Students of
English as a Foreign Language at the Institute of Teacher Training and
Education IKIP MALANG, unpublished dissertation, the University of Iowa,
Iowa City.
Mahmud, 1996. Measuring Students Performance in Reading Comprehension by
Using Multiple Choice Test Format and Short Answer Test Format. Unpublished
undergraduate Thesis, FPBS, IKIP MALANG.
Masduki, 1999. The Effectiveness of Individualized Instruction on the Students
Vocabulary Gain at the English Department Muhammadyah University of
Malang, unpublished thesis, Graduate School, IKIP MALANG.
24
Mukminatin, Nur, 1997. The Differences of Students Writing Achievements across
Different Course Levels, unpublished dissertation, Graduate School, IKIP
MALANG.
Rohimah, Umi. 1992. The Relationship between Students Pragraph Knowledge and
their Ability to write Paragraph. Unpublished undergraduate Thesis, FPBS, IKIP
MALANG.
Suharmanto, 2000. Reading Strategies in Indonesian and in English: A Comparative
Study, unpublished thesis, Graduate School, IKIP MALANG.
Sulistyaningtyas, Anis. 1996. The Correlation between Students Experience in English
Songs and their English Pronunciation. Unpublished ndergraduate Thesis, FPBS,
IKIP MALANG.
Surjosoeseno, Tjahjaning, T. 1991. The Effect of Different Types of Questions on
Reading Ability of Students at Widya Mandala Catholic University Surabaya.
unpublished thesis, Graduate School, IKIP MALANG.
Telaumbanua, Sadelie, 1992. Pengaruh Strategi Mapping terhadap Kemampuan
Menulis. unpublished thesis, Graduate School, IKIP MALANG.
Tuckman, Bruce, W. 1999 Conducting Educational Research, fifth Edition. Belmont:
Wadsworth Group.
Vockell, Edward, L., Asher, J.W. 1995. Educational Research, Second edition.
Englewood Cliffs: Prentice-Hall, Inc.
Waskito, S. Budi, 1994. The Effectiveness of Bahasa Inggris 2a untuk SMP Compared
with the English for Beginners at SMP Negeri I Kartasura, unpublished thesis,
Graduate School, IKIP MALANG.
Tidak ada komentar:
Posting Komentar